JP3046226B2 - Bit stream arrangement / restoration method for digital video tape recorder and data compression apparatus and decompression apparatus suitable for the method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an SD-VCR (Stan
(Dard Definition Videocassette Recorder), and more particularly, to the arrangement of bit streams generated as a result of variable-length coding, and more specifically, to bit streams generated as a result of variable-length coding of macroblocks included in a segment. When the code amount of the system exceeds the allowable capacity of the segment, the improved bit stream that guarantees the reproduction quality by maximally storing the excess bit amount within a range that maintains compatibility with the existing SD-VCR. The present invention relates to a method of arranging / restoring data and a data compression device and a decompression device suitable for the method.

[0002]

2. Description of the Related Art With the development of digital signal processing technology, most information transmission media have been converted to digital systems due to the advantages of digital systems over analog systems. Therefore, a digital VCR capable of maintaining high image quality even during multi-dubbing has been realized.

In a digital VCR, DCT blocks of an input image are classified into four classes according to the complexity, and bits are generated from each screen while improving the coding performance of the system by adaptively allocating bits to each class. The data volume is kept constant. Therefore, in the video compression method in the digital VCR, one screen is divided into sub-screens (segments) of a fixed size, each segment is independently encoded, and the amount of data generated for each segment is kept constant. Add.

In a conventional digital VCR, data is reproduced within a range that does not degrade the quality of a reproduced image below a certain level.
After the data compression, the remaining data is discarded in order from the data having the least effect on the image quality of the reproduced video so that the data amount per segment is kept constant. If the complexity of the input video is large, the amount of data generated will increase, but since the amount of data per segment is fixed,
In the conventional digital VCR, there is a problem that the quality of the reproduced video is reduced as the complexity of the input video is increased.

[0005] A conventional technique will be described below with reference to the drawings. The apparatus shown in FIG. 1 includes an A / D converter 10 and a data compression unit 1.
2, an error correction encoding unit 14, a channel encoding unit 16,
Reproduction unit 18, error correction decoding unit 20, data decompression unit 2
2, and a D / A converter 24. The input signal is provided to a data compression unit 12 after passing through an A / D converter 10. The audio signal is recorded without compression as a PCM signal. The image digitally converted by the A / D converter 10 has a 4: 2: 2 sampling structure according to CCIR recommendation 601. The 4: 2: 2 image has a data rate of about 167 Mbps based on the effective pixels.

[0006] The data compression unit 12 performs in-picture compression using DCT and VLC, and outputs compression-encoded video data having a data rate of 24.948 Mbps. The error correction encoding unit 14 compresses the compressed video data, audio data,
An error correction code is added to each of the subcode and auxiliary data. The data to which the error correction code has been added is recorded and encoded by the channel encoding unit 16 and recorded on a magnetic tape. The error rate coding and channel coding increase the recording data rate to 41.85 Mbps.

At the time of reproduction, after an error generated at the time of recording and reproduction is corrected by an error correction decoding unit 20 with the data reproduced through the reproduction unit 18, the data is expanded by a data decompression unit 22 and D / A conversion. The video and audio signals are restored through the unit 24.
Table 1 shows the specifications of the SD-VCR.

[0008]

[Table 1]

FIG. 2 is a diagram showing a track structure of a conventional SD-VCR. The track of the SD-VCR starts from the track start part where the head starts scanning and inserts the ITI
nd Track Information) Sector-200, Audio Sector-202, Video Sector-204, and Sub-Coordinate
The sectors are divided into the order of sectors 206, and gaps G1, G2, G3 exist between the sectors.

The ITI sector 200 has a preamble area 208, a start sync block area 210, a track information area 212, and a postamble area 214. Audio sector 202 is preamble area 2
16, 14 audio data sync block areas 2
18 and a postamble area 220.

The video sector 204 has a preamble area 222 and 149 video data sync block areas 2
24, and a postamble area 226. The subcode sector 206 includes a preamble area 228, a subcode area 230, and a postamble area 232,
And an overwrite margin (OverwriteMargin) 23
4

The head runs obliquely from the lower end of the track to the upper end. When the tape runs at a speed of 18.831mm / sec,
The inclination of the truck is 9.1668 °. One frame of a 525/60 Hz video signal is recorded on ten tracks as shown in FIG. 3A, and one frame of a 625/50 Hz video signal is recorded as shown in FIG. 3B. It is recorded on 12 tracks.

The ITI sector 200 provides track information, data position information for editing, and the like. This information is detected and a certain time interval is calculated to detect the position of the image, sound and sub-code. The sub-code records auxiliary data such as recording date and time. The subcode is read in synchronization with the video and audio recorded on the tape. Captions and captions can be recorded. Sub-co
A high-speed search of 200 times speed is also possible using a part of the code.

Azimuth recording is performed on a recording track to prevent mutual disturbance between adjacent recording tracks. The azimuth angle is ± 20 °. The required capacity of the input video signal of 4: 1: 1 or 4: 2: 0 is about five times as large as the recording capacity of the SD-VCR. Therefore, there is a need for signal compression and decompression techniques. The video compression technique used for a VCR should reflect the nature of the tape that is the recording medium. Editing function, trick play function and minimum error radio must be guaranteed. Since the tape is a successively approaching medium, independent coding should be performed on a small-scale image unit even within one screen in order to satisfy the above function. In the current standard, a segment is one independent compression coding unit. One segment is composed of five macro blocks. In the decoding process, at least one macroblock can be independently decoded. As a result, error radio waves can be reduced, which is advantageous during trick play.

FIG. 4 is a diagram showing the configuration of the data compression section shown in FIG. The data compression unit shown in FIG. 4 includes a frame memory 30, a DCT conversion unit 32, a momentum detection and a DC
T-mode determination unit 34, quantization unit 36, bit amount control unit 3
8, a variable length coding unit 40, a shuffling unit 42, a bit stream arranging unit 44, and a deshuffling unit 46.

The SD-VCR uses a compression method based on DCT and VLC. The video signal input to the frame memory 30 is a 4: 2: 2 or 4: 2: 0 video signal. The luminance (Y) signal is sampled at 13.5 MHz, and the color difference (C _r , C _b ) signal is sampled at 6.75 MHz. In the case of a 525/60 Hz signal, the color difference signal is subsampled by 2: 1 in the horizontal direction to make 4: 2: 2,
For the 625/50 Hz signal, the two color difference signals are reduced in line by 1/2 in the vertical direction so that the scanning line is reduced to と す る to be a 4: 2: 0 signal.

FIG. 5 is a diagram showing a process of setting a video signal to 4: 2: 0 according to CCIR recommendation proposal 601. 525 / 60Hz (6
25/50 Hz), the magnitude of the Y signal is 720 (720) pixels horizontally and 480 (576) lines vertically. The magnitudes of the two color difference signals are 180 (360) pixels horizontally and 480 (288) lines vertically. One DCT block is composed of 8 × 8 pixels. The color difference signal of 525 / 60Hz signal is 4 from the right end of each line.
Since an imperfect block having a size of × 8 (） size) is generated, one DCT block is formed using the upper and lower two blocks.

Therefore, the 525/60 Hz (625/50 Hz) Y signal is composed of 90 (90) horizontal DCT blocks and 60 (72) vertical DCT blocks. The two color difference signals are horizontal 22.5 (45)
And 60 (36) vertical DCT blocks. 6 (A), 6 (B) and 7 (A),
(B) shows. The macro blocks are shown in FIGS.
C _r of one by one and four Y blocks indicating the same position in the screen, as shown (B), the composed of C _b block. A macroblock is a basic unit for determining shuffling and quantization step values.

FIG. 9 is a diagram showing a shuffling method for a 525/60 Hz signal. There are five vertical regions from A to E, and each region is divided into ten small regions. Each small area is called a super block. Each superblock is composed of 27 macroblocks. Therefore, one frame screen is composed of 50 superblocks composed of 10 vertically and 5 horizontally. The 625/50 Hz signal is composed of 60 superblocks composed of 12 vertical and 5 horizontal signals.

The super blocks included in the horizontal area are recorded sequentially on the same track on the tape later. As shown in FIG. 9, one frame screen is equally divided in the vertical direction into five, and one macroblock is taken from each of the vertical regions. The order of fetching macroblocks from each area is C, B,
D, A, and E. This order is determined by the visual importance of each vertical region. As many segments as the number of macroblocks included in the superblock are generated. Since one segment is composed of screen segments distributed at many positions, there is a high probability that the amount of information will be similar for each segment. The more the information amount is constant, the more efficient the code amount can be fixed. Such mixing of the entire screen in macroblock units is called shuffling. Compression encoding is performed completely independently on a segment basis, and generates the same amount of code for every segment.

Each macroblock is shuffled when a segment is formed, but is returned to the original position immediately before recording.
This is called deshuffling. The reason for performing the deshuffling is to construct a visually good screen so that data read during high-speed reproduction is arranged as continuously as possible. The playback screen at the time of high-speed playback is a mosaic of small screens distributed over many tracks. By fixing the code amount for each segment, it is possible to prevent the image quality during trick play from lowering to the maximum.

In FIG. 10, five super blocks in the horizontal direction are recorded on one track. Since there are ten lines in the vertical direction, one screen is consequently recorded on ten tracks. 13 for video recording in one track
The five sync blocks are in one-to-one correspondence with 135 macro blocks brought from the five super blocks. S
In the D-VCR, the size of the DCT block differs depending on the motion information. When there is almost no movement, an 8x8 size D
Take CT. On the other hand, if it is determined that the motion is large, the sum and difference of two vertically adjacent pixels in the block are calculated, and two 4 × 8 DCTs are obtained. FIG. 11A shows 8 × 8D.
FIG. 11B shows a 2 × 4 CT block.
Indicates a × 8 DCT block.

Each DCT block is multiplied by a corresponding weight matrix to reduce the range value differently for each coefficient. The weight matrix has a smaller value as the frequency component is higher. Each DCT block is classified into one of four classes. Table 2 shows the class definitions.

[0024]

[Table 2]

The AC component in the DCT block is divided into four areas 0 to 3, and coefficients in the same area are quantized by the same quantization step. The elements that determine the quantization step are the class and the area within the DCT block. Each quantization system is 1, 1/2, 1/4,. . Is one of the powers of two. FIG. 12A shows 8 × 8.
FIG. 12 shows the area division of the DCT block.
(B) shows the area division of the 2 × 4 × 8 DCT block.

The quantized DCT coefficients are one-dimensionally rearranged from a low frequency to a high frequency. The DC coefficient is encoded as 9 bits, and the AC coefficient is encoded as a two-dimensional Huffman code according to the size. One segment has 30 DCT blocks, and 30 recording areas are separately prepared. All recording areas take the Huffman code generated from their own DCT block with the highest priority. The remaining bits are recorded in the recording area of another DCT block in the same macro block. When the recording in the macro block is completed, bits are given or received according to the remaining or insufficient state between the five macro blocks, and the recording is appropriately performed.

Even if the Huffman code of 30 blocks is finally obtained in one segment, the remaining recording area cannot be reused even though the recording area remains. Conversely, the bit amount remaining when the recording area is insufficient is discarded. Thus, each segment is encoded independently. As described above, the conventional SD-VCR reduces the bit amount by changing the quantization step within a range where a predetermined level of image quality is guaranteed if the bit amount of the currently coded segment is too large. Discarding the amount degrades the reproduction quality of a screen with a large amount of bits generated like a screen with a high degree of complexity.

[0028]

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has been made by maximizing the use of a bit stream generated as a result of variable length coding. An object of the present invention is to provide a new bit stream arranging method for maintaining the quality of a reproduced image.

Another object of the present invention is to provide a bit stream restoring method suitable for the bit stream arranging method. It is still another object of the present invention to provide a data compression apparatus suitable for the method of arranging the bit stream. Still another object of the present invention is to provide a data decompression device suitable for the above-mentioned restoration method.

[0030]

According to the present invention, there is provided a method for arranging a bit stream, comprising a plurality of macroblocks, wherein the macroblocks comprise a plurality of D blocks.
SD in which a bit stream generated as a result of quantizing and variable-length encoding DCT data of each DCT block is arranged in a fixed-length segment including a CT block.
-In the VCR bit stream arrangement method, (a)
Arranging a bit stream of each DCT block generated as a result of quantizing and variable-length coding DCT data of each DCT block in a DCT block in the macro block; B) rearranging the bit amount remaining in the macroblock to another DCT block; and c) rearranging the bit amount remaining in the step b) between macroblocks in the segment.
(D) relocating the amount of bits remaining in the step (c) to another segment having a macroblock of the same content as that of the previous or subsequent frame.

According to another aspect of the present invention, there is provided a method for restoring a bit stream, comprising the steps of:
A method for restoring a bit stream generated as a result of quantization and variable length coding of DCT data of a DCT block from a segment arranged hierarchically at a macroblock or segment level, comprising: Determining whether the macroblock included in the segment has an incomplete bitstream;
If it is determined in the process that there is a macroblock having an incomplete bitstream, a process of restoring an extra bitstream from another macroblock in the segment, and (c) a process of the above (b) If the macroblock having the incomplete bitstream is not completely restored, a search is performed in the search area for a segment including the error-processed macroblock; Restoring an incomplete bit stream using a bit stream included in the error-processed macroblock in the segment.

A data compression apparatus according to the present invention for achieving the above-mentioned object is a frame memory, a DCT.
A DCT block comprising a conversion unit, a momentum detection and DCT mode determination unit, a quantization unit, a bit amount control unit, a variable length coding unit, a shuffling unit, a bit stream arranging unit, and a deshuffling unit. , Macroblocks, and DCTs from segments hierarchically arranged at the segment level
Search for previous or subsequent frames in an SD-VCR data compression device suitable for arranging bit streams generated as a result of quantization and variable length coding of DCT data of blocks. A macroblock search unit that searches for macroblocks having the same contents in the area;
Calculating a position difference between the macroblock searched by the macroblock searching unit and the currently processed macroblock, setting an STA value corresponding thereto and providing the STA value to the bitstream arranging unit; Parts.

According to another aspect of the present invention, there is provided a data decompression apparatus for quantizing DCT data of a DCT block from a segment arranged hierarchically at a DCT block, macro block, or segment level. And an SD-VCR data decompression device suitable for restoring a bit stream generated as a result of variable-length encoding,
A shuffling unit for reconstructing a segment from the reproduced data; and a macroblock in which variable-length coded data and errors have been generated in macroblock units from the segment output from the shuffling unit. A bitstream reverse arranging unit for extracting STA information, and variable length coded data in macroblock units output from the bitstream reverse arranging unit, are input, and a variable A long decoding unit, and an inverse DCT for performing an inverse DCT transform on the variable length decoded data provided from the variable length decoding unit to restore and output a video signal in a time domain.
A de-shuffling unit for de-shuffling the data in units of macroblocks that have been subjected to the inverse DCT transformation through the inverse DCT unit, returning the data to the original position and outputting the data, and an output from the bit stream inverse arrangement unit An concealment processing unit that receives STA information of a macroblock in which an error has occurred and generates position information of a macroblock corresponding to the macroblock in which an error has occurred in a previous or subsequent frame; It is determined whether an error has occurred in a macroblock provided from the deshuffling unit based on information provided from the concealment processing unit. If an error has occurred, the macroblock corresponding to the previous or subsequent frame is determined. A frame memory for substituting and outputting a macroblock in which an error has occurred.

[0034]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to FIG. FIG. 13 shows a bit stream according to the present invention.
6 is a flowchart illustrating a method for arranging programs. As shown in FIG. 13, the bit stream arranging method according to the present invention includes a DCT block arranging process (operation 1200), a macro block rearrangement operation (operation 1210), a segment allocation operation (operation 1220), and a segment allocating process. (1230).

At step 1200, the DC of each DCT block is
A bit stream of each DCT block generated as a result of quantization and variable length coding of the T data is arranged in a DCT block in the macro block. The above
The recording capacity of the DCT block in the macro block is fixed.
Yes, the amount of bits exceeding the recording capacity of the DCT block
Is left unrecorded. 1210 In step rearranges the remaining Ru bit amount without being recorded in operation 1200 to the other DCT blocks in the macroblock. In the 1220 stage, 12
10 step to rearrange the bits remaining without being relocated to another DCT block to another macro block within the segment. At step 1230, another macro block is set at step 1220.
The remaining bits that have not been relocated to the previous frame are relocated to another segment having a macroblock of the same content as the previous or subsequent frame.

FIG. 14 is a flowchart specifically illustrating the step 1230 shown in FIG. The inter-segment rearrangement process (operation 1230) shown in FIG. 14 includes a macroblock search operation (operation 1300), a macroblock error processing operation (operation 1310), and a rearrangement operation (operation 1320). In operation 1300, a macroblock having the same content as the macroblock existing in the search space of the frame to be encoded is searched in the search space of the previous or subsequent frame. In operation 1310, the recording area corresponding to the searched macro block is subjected to error processing in a segment including the macro block searched from operation 1300. 1310 at 1320
The bits remaining in step 1310 are rearranged in the error-processed recording area through steps.

FIG. 15 is a diagram showing a bit stream arrangement method according to the present invention shown in FIG. In FIG. 15, the first segment 140 and the second segment 150 are each composed of five macroblocks. here,
The first segment 140 is the segment having the macroblock 142 in which the excess bits have been generated, and the second segment 150 is the segment having the searched macroblock 152.

The amount of bits exceeded in macroblock 142 is recorded in searched macroblock 152. The macro block 142 in which the excess bit amount has occurred does not have an EOB (End of Block) code, thereby notifying that the excess bit amount exists, and the searched macro block 152 has a macro with a special STA code. The position information of the block 142 is recorded.

At the time of restoration, the data expansion unit (not shown) uses EO.
If a macroblock having no B is generated, a macroblock having a special STA code is searched for in the preceding and following segments. The excess bits are taken from the searched macroblock to restore the complete data. On the other hand, the data of the searched macroblock is restored using the macroblock at the same position in the previous or subsequent frame.

In the present invention, if there are bits remaining in the currently coded segment, the same or similar macroblock is searched for in the search area of the previous or subsequent frame area, and the bits remaining in the searched macroblock are searched. Is placed.
In this case, the STA information of the macro block is used to normally restore the contents to be recorded in the searched macro block.

FIG. 16 is a diagram showing the data arrangement of macro blocks in a segment. In FIG. 16, the macro block includes a quantization number area 150, an STA area 152, a first
Luminance block area 154, second luminance block area 15
6, a third luminance block area 158, a fourth luminance block area 160, a first chrominance block area 162, and a second chrominance block area 164.

Each of the quantization number area 150 and the STA area has 4 bits. First luminance block area 154, second luminance block area 156, third luminance block area 15
The eighth and fourth luminance block areas 160 each have a recording capacity of 14 bytes. The first color difference block area 16
The second and second color difference block areas 164 each have a recording capacity of 10 bytes. As a result, each macroblock is 77
It has a recording capacity of bytes. The STA code is a 4-bit code (s3, s2, s1, s) provided for error concealment.
0). Table 3 shows STA codes specified in the standard SD-VCR.

[0043]

[Table 3]

According to the present invention, instead of recording the amount of bits remaining in the searched macroblock, error processing is performed using STA information, and when restoring, a macroblock of the frame before or after having the same contents is restored. Put it on SD-VCR
And keep it interchangeable. At this time, the search space for searching for a macroblock having the same content between the previous and next frames is limited to the space covered by the superblock. As described in the description of the shuffling method with reference to FIG. 11, adjacent macroblocks in the same superblock are shuffled in adjacent segments, so that a macroblock to be matched at the time of decoding is smoothly searched. Thus, the search space is limited within the superblock.

The bit stream arranging method according to the present invention employs "0010" or "010" in the first embodiment.
An STA code having a value of “0” is used.
If a macroblock having a STA code having a value of “010” or “0100” is generated, a macroblock to be matched among segments existing within a predetermined search range based on the position, that is, having a remaining bit amount. In order to search for a power macroblock, the search range can be limited to, for example, ± 8 to improve the efficiency of the search algorithm.

In the second embodiment of the bit stream arranging method according to the present invention, the position information of the macro block matched to the area where the DC coefficient is recorded in the DCT block is recorded. That is, at the time of decryption, "00"
When a macroblock having an STA code having a value of "10" or "0100" is generated, an area where the DC coefficient of each DCT block is recorded is read to search for a matching macroblock.

The first and second embodiments use the conventional SD-V
Ensures interchangeability with CR. However, since an error may actually occur in a macroblock, a method of recording a special identification code in the quantization number information (QNO) for this purpose may be considered. In the third embodiment of the bit stream arrangement method according to the present invention, the position of a macroblock to be matched is designated using STA information. That is, in Table 3, “0010” and “010”
A segment in which a macroblock to be matched is located is designated using STA information including direction information bits and offset information bits of a reserved area other than 0 ".

FIG. 17 shows ST in the third embodiment of the present invention.
6 is a diagram illustrating a method of configuring an A code. The bits constituting the STA code are s3, s2,
When s1 and s0 are used, the s3 bit is used as direction information bits indicating forward and backward directions, and the s2 to s0 bits are used as position information. In the third embodiment, the exchangeability with the conventional SD-VCR is guaranteed, but there are some problems in consideration of future standard expansion.

FIG. 18 is a block diagram showing the configuration of a data compression unit suitable for the bit stream allocation method according to the present invention. The device shown in FIG.
CT conversion section 172, momentum detection and DCT mode determination section 174, quantization section 176, bit amount control section 178, variable length coding section 180, shuffling section 182, bit stream arrangement section 184, deshuffling Section 186, macroblock search section 188, and STA value setting section 190
And

The macroblock search unit 188 searches for a macroblock having the same contents in the search area of the previous or subsequent frame. The STA value setting unit 190 calculates a position difference between the macroblock searched by the macroblock searching unit 188 and the currently processed macroblock, sets an STA value corresponding to the difference, and sets a bit stream.
Provided to the system arrangement unit 184.

If there is a bit amount remaining after performing the bit stream arrangement in segment units, the bit stream arranging section 184 assigns a corresponding macro block of the segment including the macro block searched by the macro block searching section 188 to the corresponding macro block. The remaining bit amount is allocated, and the STA code of the corresponding macroblock is recorded as the value set by the STA value setting unit 190.

FIG. 19 is a block diagram showing a data decompression unit corresponding to the data compression unit shown in FIG. FIG.
Is composed of a shuffling unit 1800, a bitstream reverse arrangement unit 1802, a variable length decoding unit 1804, an inverse DCT unit 1806, a deshuffling unit 1808, a frame memory 1810, and a concealment processing unit 1812. Is done.

The shuffling unit 1800 reconstructs a segment from data reproduced through a reproducing unit (not shown) and provides the reconstructed segment to the bitstream reverse arrangement unit 1802. The bit stream reverse arrangement section 1802 is provided in the shuffling section 18.
The macroblock information is extracted from the segment output from 00 and provided to the variable length decoding unit 1804, and the STA information of each macroblock is output.

The variable length decoding section 1804 performs variable length decoding on a macroblock basis, and outputs the result to the inverse DCT section 18.
06. An inverse DCT unit 1806 performs an inverse DCT transform on the variable-length decoded data provided from the variable length decoding unit 1804 to restore a time-domain video signal and provides it to a deshuffling unit 1808. .

The deshuffling unit 1808 returns the data in units of macroblocks subjected to inverse DCT conversion to the original position and provides the data to the frame memory 1810. Concealment processing unit 18
Numeral 12 receives STA information of a macroblock output from the bitstream reverse arrangement unit 1802, and generates position information of a macroblock corresponding to a macroblock in which an error has occurred in a previous or subsequent frame. .

The frame memory 1810 includes the concealment processing unit 18
The deshuffling unit 18 is provided by the information provided from the
It is determined whether an error has occurred in the macroblock provided from step 08. If an error occurs, the macroblock in which the error has occurred is substituted for the corresponding macroblock of the previous or subsequent frame. Let it. Frame memory 1
Reference numeral 810 outputs the error-corrected video data.

The bitstream reverse arrangement unit 1802 determines whether there is a bit amount that cannot be restored in a macroblock among the segments provided from the shuffling unit 1800. Whether or not the amount of bits that cannot be restored is determined by checking whether the recording space of each macroblock has been completed with the EOB code. If the processing is terminated by the EOB code, it is recognized that the data of the corresponding macroblock has been completely restored. Otherwise, it is recognized that information remaining in another segment has been recorded.

According to the first restoring method corresponding to the first embodiment of the bit stream arranging method according to the present invention, if an EOB code is not found in a macroblock of a segment, "0010" of the segments within the search range is used. "Or" 010
A search is made for a macroblock in which a STA code having a value of 0 "is recorded. The missing bits are restored using data recorded in the searched macroblock. Because it does not have its own variable length coded data,
This should be restored using the data of the macroblock at the corresponding position in the system. The bitstream reverse arrangement unit 1802 conceals the segment number including the searched macroblock and the number of the macroblock, and conceals the processing unit 1812
To provide.

According to the second restoring method according to the second embodiment of the bit stream arranging method of the present invention, if no EOB code is found in the macroblock of the segment, "0010" of the segments within the search range is used. "Or" 01
A macro block in which an STA code having a value of "00" is recorded is searched. Using the position information recorded in the DC coefficient recording area of the searched macro block, a remaining bit amount is matched with the generated macro block. If it is determined that the macroblock is a matching macroblock, the missing bit amount is restored using the data recorded in the area where the AC coefficient is recorded.

According to the third restoring method corresponding to the third embodiment of the bit stream arranging method of the present invention, if an EOB code is not found in a macroblock of a segment, the segment in the search range is reserved. A macro block in which the STA code is recorded is searched. Using the STA information of the searched macro block, it is checked whether the remaining bit amount is a macro block matched with the generated macro block. If it is determined that the macroblock is a macroblock to be matched, the missing bit amount is restored using data recorded in the area where the AC coefficient is recorded.

[0061]

As described above, in the bit stream arranging method according to the present invention, if a bit amount remaining in a segment unit is generated, the bit amount is propagated to an adjacent segment and recorded, thereby deteriorating the reproduced image quality. To prevent
According to the method of arranging the bit stream according to the present invention, a macro block having the same contents between the previous and next frames is selected as a macro block for recording the remaining bit amount. Absent.

The method of arranging the bit stream according to the present invention uses the STA code to specify the macro block when recording the excess bit amount.
The exchangeability of R is maintained. The first and second embodiments of the bit stream arranging method according to the present invention maintain interchangeability with the conventional SD-VCR by using the predefined STA information of the STA information of the macroblock. Add.

The third embodiment of the bit stream arranging method according to the present invention uses the reserved STA information among the STA information of the macro block to realize the conventional SD-stream.
Makes it compatible with VCRs.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an overall configuration of an SD-VCR.

FIG. 2 is a diagram showing a track structure of an SD-VCR.

3A is a diagram showing a track structure for recording a 525/60 Hz signal, and FIG. 3B is a diagram showing a track structure for recording a 625/50 Hz signal.

FIG. 4 is a block diagram illustrating a detailed configuration of a data compression unit illustrated in FIG. 1;

FIG. 5 shows a video signal according to CCIR recommendation proposal 601 at 4:
It is a figure which shows the process of making it 2: 0.

FIG. 6 is a diagram showing a frame configuration of a 4: 2: 2 signal on a macroblock basis.

FIG. 7 is a diagram showing a frame configuration of a 4: 2: 0 signal on a macroblock basis.

FIGS. 8A and 8B show 4: 2: 2 signals and 4: 2: 0
FIG. 3 is a diagram illustrating a macroblock of a signal.

FIG. 9 is a diagram showing a shuffling method using a 525/60 Hz signal.

FIG. 10 is a diagram showing area division for a segment configuration.

FIG. 11A shows an 8 × 8 DCT block,
(B) shows a 2 × 4 × 8 DCT block.

FIG. 12A shows an area division of an 8 × 8 DCT block, and FIG. 12B shows an area division of a 2 × 4 × 8 DCT block.

FIG. 13 is a flowchart showing a bit stream arrangement method according to the present invention.

FIG. 14 is a flowchart more specifically showing a rearrangement process between segments shown in FIG. 13;

FIG. 15 is a diagram illustrating a bit stream arrangement method according to the present invention shown in FIG. 13;

FIG. 16 is a diagram showing a data arrangement of macro blocks in a segment.

FIG. 17 is a diagram illustrating a method of configuring an STA code.

FIG. 18 is a block diagram showing a configuration of a data compression unit according to the present invention.

FIG. 19 is a block diagram showing a configuration of a data decompression unit according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 A / D converter 12 Data compression part 14 Error correction encoding part 16 Channel encoding part 18 Reproduction part 20 Error correction decoding part 22 Data decompression part 24 D / A converter 30 Frame memory 32,172 DCT transformation unit 34,174 Momentum detection and DCT mode determination unit 36,176 Quantization unit 40,180 Variable length encoding unit 42,182,1800 Shuffling unit 44,184 Bit stream arrangement unit 46 , 186, 1808 Deshuffling unit 170, 1810 Memory 178 Bit amount control unit 190 STA value setting unit 188 Macro block search unit 1802 VLC bitstream reverse arrangement unit 1804 Variable length decoding unit 1806 Inverse DCT unit 1812 Concealment processing unit

Claims (13)

(57) [Claims] 1. A macroblock comprising a plurality of macroblocks, said macroblocks being generated as a result of quantizing and variable-length coding DCT data of each DCT block into fixed-length segments including a plurality of DCT blocks. (A) a bit stream of each DCT block generated as a result of quantizing and variable-length coding DCT data of each DCT block; Stream in the macroblock
And (b) predetermined DCT block data in the step (a).
The amount of bits in excess of the amount
Relocating to a block; and (c) the excess not relocated in step (b).
A process of rearranging the bit amount between macroblocks in segment, the excess is not rearranged; (d) (c) Process
Amount of bits previously, or subsequent frame - beam arrangement method - the bit string including the steps to be relocated to another segment having the macro block of the arm and the same content. 2. In the step (d), a macroblock having the same contents as a macroblock existing in a search space of a frame to be encoded is searched in a search space of a previous or subsequent frame. Error processing a recording area corresponding to the searched macro block with a segment including the macro block searched in the macro block searching step; Redistribution in the process
Re-arranging the excess bit amount that has not been set.
Method. 3. The bit stream arrangement method according to claim 2, wherein the size of the search space is equal to or smaller than a super block. 4. The method according to claim 2, wherein the error processing process uses STA information. 5. In the error processing, the STA information is “0”.
The bit stream arrangement method according to claim 4, wherein the bit stream is "010" or "1100". 6. The position information of a segment in which a bit amount remaining in an area for recording a DC coefficient of a macroblock in which the excess bit amount is not rearranged and which has been rearranged is generated. Item 4. The bitstream arrangement method according to Item 4. 7. The error processing is performed by setting the STA signal to "0".
000 "," 0010 "," 1010 "," 010 "
0 "," 1100 "," 0110 "," 1110 ",
5. The method according to claim 4, wherein a bit stream other than "0111" is used. 8. At least one bit among bits constituting the STA signal is used as a direction information bit indicating a previous or a subsequent frame, and the remaining bits are used from a current segment to a segment to be rearranged. 8. The bit stream arrangement method according to claim 7, wherein the bit stream is used as position information bits indicating an offset of the bit stream. 9. The bits constituting the STA signal are set to “S0,
9. The bit stream arrangement method according to claim 8, wherein when S1, S2, and S3 "are used, the S0 bit is used as direction information bits, and the S1, S2, and S3 bits are used as position information bits. 10. A bit stream generated as a result of quantization and variable-length coding of DCT data of a DCT block from segments hierarchically arranged at DCT block, macro block, and segment levels. (A) determining whether there is a macroblock having an incomplete bit stream in the segment; and (b) macro having an incomplete bit stream in the (a) step. If it is determined that a block exists, an extra bit stream is stored from other macroblocks in the segment.
(C) if the macroblock having the incomplete bit stream is not completely restored in the step (b), the error-processed macroblock in the search area is deleted. (D) using the bit stream included in the error-processed macroblock in the segment searched in step (c) to detect an incomplete bit stream. And restoring the bit stream. 11. The method according to claim 10, wherein the search area is equal to or smaller than a super block. 12. A frame memory, a DCT transforming section, a momentum detecting and DCT mode determining section, a quantizing section, a bit amount controlling section, a variable length coding section, a shuffling section, a bit stream arranging section, A DCT block, a DCT block, a macro block, and a DCT block of a DCT block from a segment hierarchically arranged at a segment level.
In an SD-VCR data compression apparatus for compressing a bit stream generated as a result of quantization and variable length coding of data, the same data is used within a search area of a previous or subsequent frame. A macroblock search unit for searching for a macroblock having contents, calculating a position difference between the macroblock searched by the macroblock search unit and the currently processed macroblock, and setting an STA value corresponding thereto. A bit stream arranging unit for providing the bit stream arranging unit with an STA value, wherein the bit stream arranging unit has a bit amount remaining after performing the bit stream arranging in segment units;
The bit amount remaining in the corresponding macroblock of the segment including the macroblock searched by the macroblock search unit is arranged, and the STA code of the macroblock is recorded as the value set by the STA value setting unit. SD-VCR data compression device. 13. A bit stream generated as a result of quantization and variable-length coding of DCT data of a DCT block from segments hierarchically arranged at DCT block, macro block, and segment levels. A shuffling unit for reconstructing a segment from the reproduced data; and a variable-length encoded data in macroblock units from the segment output by the shuffling unit. A bitstream reverse arrangement unit for extracting STA information of a macroblock in which a data and an error have occurred, and variable-length coded data in macroblock units output from the bitstream reverse arrangement unit. And a variable-length decoding unit for performing variable-length coding, and variable-length-decoded data provided from the variable-length decoding unit. An inverse DCT unit for performing an inverse DCT transform to restore and output a video signal in a time domain, and de-shuffling the inverse DCT transformed data in units of macroblocks through the inverse DCT unit to an original position. A deshuffling unit for outputting the STA information of the macroblock in which the error is generated, which is output from the bitstream reverse arrangement unit, and outputs an error in a previous or subsequent frame. A concealment processing unit for generating position information of a macroblock corresponding to the selected macroblock, and whether an error has occurred in a macroblock provided from a deshuffling unit using information provided from the concealment processing unit. Is determined, and if an error occurs, the macroblock in which the error occurred is substituted for the macroblock corresponding to the previous or subsequent frame and output. A data decompressor for an SD-VCR including a frame memory.